论文部分内容阅读
The paper describes an efficient direct method to solve an equation Ax = b, where A is a sparse matrix, on the Intel®Xeon PhiTM coprocessor. The main challenge for such a system is how to engage all available threads (about 240) and how to reduce